PauseAI US

PauseAI US is the United States national chapter of the PauseAI international federation, organized as an independent 501(c)(3) nonprofit headquartered in San Francisco, California. The organization runs a grassroots campaign for a global, treaty-backed pause on the development of the most powerful frontier AI systems until representative democratic oversight can be established and development can proceed safely. Holly Elmore, who holds a PhD in evolutionary biology from Harvard and previously researched wild animal welfare, founded the U.S. chapter in 2023 after the March 2023 Future of Life Institute open letter calling for a pause on AI systems more powerful than GPT-4. Elmore organized the first group AI safety protest in the United States — approximately 25 people outside OpenAI's San Francisco headquarters in September 2023 — and built the national organization from that starting point.

Structure and relationship to PauseAI Global

PauseAI US operates semi-autonomously within the global PauseAI federation. Unlike most national chapters, which are volunteer-run, PauseAI US maintains its own paid staff and independent donor base, and officially incorporated as a 501(c)(3) nonprofit in 2024. As of mid-2026 the paid core team comprises Elmore as Founder and Executive Director, Felix DeSimone as Organizing Director (responsible for chapter support and volunteer training), and Alvaro Cuba as Development Director. An earlier staff roster included Lee Green as Director of Operations; the organization's prior EA Forum AMA described Lee Green's background as more than twenty years in strategy consulting.

Elmore has described herself as co-founder of the broader PauseAI movement alongside Joep Meindertsma, who built and leads PauseAI globally from the Netherlands. The U.S. chapter shares PauseAI's international protest calendar, treaty-advocacy platform, and policy proposal — but operates as a distinct legal entity with its own fundraising, legislative strategy, and local organizing infrastructure.

The U.S. chapter coordinates a network of city-level local groups that serve as its primary organizing unit. The chapter started 2024 with a single local group in San Francisco — "the belly of the beast," as the organization's own newsletter described it, owing to the density of frontier AI lab offices within walking distance — and grew to six active groups by January 2025 (San Francisco, New York City, Chicago, Portland, Phoenix, and Washington DC). The Q2 2026 donor update records 47 local groups across 29 states, with a stated goal of 75 active local groups by end of 2026.

Congressional strategy

The dominant strategic bet for PauseAI US is direct constituent lobbying: training and equipping volunteers to meet in person with their own members of Congress, under the theory that constituents speaking plainly about catastrophic AI risk carry weight that professional lobbyists do not. Elmore has articulated this as "rebalancing the center" — applying visible democratic pressure from outside the AI ecosystem to shift what policymakers consider an acceptable level of risk, contrasting with the predominantly insider-focused approach of traditional AI safety work.

By the end of 2025 the organization had conducted 192 meetings with congressional offices since its founding, covering House and Senate members across both parties. Its first lobbying effort in 2024 reached 25 congressional offices; by January 2025 this included seven in-district meetings where constituent volunteers traveled to their home districts to meet representatives directly.

The organization's largest single advocacy event to date was PauseCon DC in April 2026, a two-day Washington gathering in which volunteers from 18 states held 74 meetings with congressional offices in the span of 48 hours. By mid-2026 PauseAI US was also pursuing a bipartisan, bicameral letter to Secretary of State Rubio calling on the United States to lead negotiations on an international AI treaty to prevent superintelligence, with 15 congressional offices reporting strong interest in signing.

The organization's federal legislative priorities include the Ban Superintelligence Act, the Bipartisan Kill Switch Act, and the AI Risk Evaluation Act.

Protest and public organizing

Street protest remains a core tactic alongside the lobbying work. The San Francisco Bay Area chapter, the originating unit of the U.S. operation, has staged repeated demonstrations outside OpenAI's Mission-campus headquarters, Anthropic's Market Street offices, and Meta's San Francisco offices. PauseAI US participates in PauseAI's international protest waves — the May 2024 day of action in thirteen countries ahead of the AI Seoul Summit and subsequent international mobilizations — while also organizing U.S.-specific actions, including the March 2026 Stop the AI Race march in San Francisco, described by Elmore as the largest AI safety protest in U.S. history. A protest sign from the chapter's 2024 OpenAI demonstration was acquired by the Museum of Protest in Washington DC.

The chapter network runs recurring public outreach through First Friday Flyering, a monthly cadence in which volunteers across multiple cities simultaneously distribute materials and recruit passersby in public spaces.

Legislative record

PauseAI US has pursued a dual track of federal and state legislative engagement. At the state level, the organization supported California's SB 53, which PauseAI US describes as the first major AI safety law focused on catastrophic risks, signed into law in 2025. Also in 2025, the U.S. Senate struck down AI pre-emption language in a federal bill 99 to 1, a vote PauseAI US counted as a significant defensive victory — preventing federal legislation from stripping states' ability to enact their own AI safety laws. At the federal level PauseAI US attended the UN Summit for the Future in 2024 alongside delegations tracking international AI governance.

Public profile and the April 2026 incident

The April 2026 attack on Sam Altman's residence brought brief and unwanted national scrutiny to PauseAI US. Reporting noted that suspect Daniel Moreno-Gama had been active on PauseAI's public Discord server; Elmore emphasized that he was never a member of their organizations and that PauseAI US "explicitly rejects" violence and illegal activity. The incident also clarified the organizational boundary between PauseAI US and Stop AI: Elmore told Fortune that she had expelled Stop AI's cofounders from PauseAI due to disagreements over tactics, specifically Stop AI's willingness to engage in civil disobedience involving law-breaking — a line PauseAI US does not cross. PauseAI US defines its activism as strictly peaceful and lawful democratic organizing.

The organization has received coverage in the New York Times, Bloomberg, Fortune, NPR, and WIRED. As of 2026 it operates with a budget in the low hundreds of thousands, funded primarily through grassroots donations with a reported median donation of $100, supplemented by SFF speculation grants; the Q2 2026 donor update describes the organization as seeking $450,000 for a full year of operations.